Soft robotics is a growing area of research which utilises the compliance and adaptability of soft structures to develop highly adaptive robotics for soft interactions. One area in which soft robotics has the ability to make significant impact is in the development of soft grippers and manipulators. With an increased requirement for automation, robotics systems are required to perform task in unstructured and not well defined environments; conditions which conventional rigid robotics are not best suited. This requires a paradigm shift in the methods and materials used to develop robots such that they can adapt to and work safely in human environments. One solution to this is soft robotics, which enables soft interactions with the surroundings whilst maintaining the ability to apply significant force. This review paper assess the current materials and methods, actuation methods and sensors which are used in the development of soft manipulators. The achievements and shortcomings of recent technology in these key areas are evaluated, and this paper concludes with a discussion on the potential impacts of soft manipulators on industry and society.

Agriculture provides an unique opportunity for the development of robotic systems; robots must be developed which can operate in harsh conditions and in highly uncertain and unknown environments. One particular challenge is performing manipulation for autonomous robotic harvesting. This paper describes recent and current work to automate the harvesting of iceberg lettuce. Unlike many other produce, iceberg is challenging to harvest as the crop is easily damaged by handling and is very hard to detect visually. A platform called Vegebot has been developed to enable the iterative development and field testing of the solution, which comprises of a vision system, custom end effector and software. To address the harvesting challenges posed by iceberg lettuce a bespoke vision and learning system has been developed which uses two integrated convolutional neural networks to achieve classification and localization. A custom end effector has been developed to allow damage free harvesting. To allow this end effector to achieve repeatable and consistent harvesting, a control method using force feedback allows detection of the ground. The system has been tested in the field, with experimental evidence gained which demonstrates the success of the vision system to localize and classify the lettuce, and the full integrated system to harvest lettuce. This study demonstrates how existing state‐of‐the art vision approaches can be applied to agricultural robotics, and mechanical systems can be developed which leverage the environmental constraints imposed in such environments.

Conditional models provide a method of achieving varying behaviors by using hybrid and passive mechanical structures.

Underwater soft robots are challenging to model and control because of their high degrees of freedom and their intricate coupling with water. In this letter, we present a method that leverages the recent development in differentiable simulation coupled with a differentiable, analytical hydrodynamic model to assist with the modeling and control of an underwater soft robot. We apply this method to Starfish, a customized soft robot design that is easy to fabricate and intuitive to manipulate. Our method starts with data obtained from the real robot and alternates between simulation and experiments. Specifically, the simulation step uses gradients from a differentiable simulator to run system identification and trajectory optimization, and the experiment step executes the optimized trajectory on the robot to collect new data to be fed into simulation. Our demonstration on Starfish shows that proper usage of gradients from a differentiable simulator not only narrows down its simulation-to-reality gap but also improves the performance of an open-loop controller in real experiments.

The development and use of architectured structures is changing the means by which we design and fabricate soft robots. These materials utilize their topology and geometry to control physical and mechanical structural properties. We propose an architectured structure based on trimmed helicoids that allows for independent regulation of the bending and axial stiffness which facilitates tuneability of the resulting soft robot properties. Leveraging FEA and computational analysis we select a geometry that provides an optimal trade-off between controllability, sensitivity to errors in control, and compliance. By combining these modular trimmed helicoid structures in conjunction with control methods, we demonstrate a meter-scale soft manipulator that shows control precision, large workspace, and compliant interactions with the environment. These properties enable the robot to perform complex tasks that leverage robot-human and robot-environment interactions such as human feeding and collaborative object manipulation.

Advances in soft sensors coupled with machine learning are enabling increasingly capable wearable systems. Since hand motion in particular can convey useful information for developing intuitive interfaces, glove-based systems can have a significant impact on many application areas. A key remaining challenge for wearables is to capture, process, and analyze data from the high-degree-of-freedom hand in real time. We propose using a commercially available conductive knit to create an unobtrusive network of resistive sensors that spans all hand joints, coupling this with an accelerometer, and deploying machine learning on a low-profile microcontroller to process and classify data. This yields a self-contained wearable device with rich sensing capabilities for hand pose and orientation, low fabrication time, and embedded activity prediction. To demonstrate its capabilities, we use it to detect static poses and dynamic gestures from American Sign Language (ASL). By pre-training a long short-term memory (LSTM) neural network and using tools to deploy it in an embedded context, the glove and an ST microcontroller can classify 12 ASL letters and 12 ASL words in real time. Using a leave-one-experiment-out cross validation methodology, networks successfully classify 96.3% of segmented examples and generate correct rolling predictions during 92.8% of real-time streaming trials.